At the very heart of the digital economy lie data centers, critically more than ever to our digitized way of life. As the world adapts to a new and hybrid mode of working, more innovations will be necessary for companies to increase sustainability. The challenge is that IT as a whole has a significant carbon footprint, which is expected to increase as businesses move towards digital acceleration.
Data centers have primarily been one of the most energy-intensive parts of an organization. In 2018, data centers worldwide consumed almost 2% of global electricity and emitted CO2 roughly of the same levels as the airline industry. With global data multiplying 2-fold every 4 years, these emissions will only increase.
Thus, going into 2022, companies are expected to go beyond the best practices of increasing data center energy efficiency. Here are some trends to expect in the future.
Building Sustainable Edge Infrastructure
Edge deployments might be smaller than big data centers, but customers will demand them to match the latter’s efficiency and resilience. Building sustainable edge at scale will require advanced management systems to check on heating and cooling in a way that increases operational efficiency. Innovative approaches will be essential to drive cooling, especially in unmanned environments.
Air cooling in urbanized and harsh locations presents issues of dust and other contaminants floating about. Even if filters are used, they represent tedious requirements of servicing and frequent replacements. Blowing around such material in remote edge centers is not feasible. Thus, liquid cooling will be essential to increase the energy efficiency of unmanned and sealed data centers. Some trends to expect here would be chassis-based immersive cooling or direct-to-chip cooling.
Introducing Sustainable Construction in Building Green Data Centers
Until now building green data centers has led to discussions and practices on reducing operational energy. This means the energy used to power, heat, or cool these facilities. These metrics can be easy to measure and track. However, this might not be enough. This is because IT hardware will continue to evolve and become efficient, and renewable energy will be extensively used by the energy grids. Thus, companies won’t be able to reduce carbon footprint, only through operational efficiencies.
The construction of data centers will also be overhauled. Two materials used in data centers are cement and steel, both most carbon-intensive construction materials. A ton of cement on average produces 1.25 tons of CO2, especially emanating from silica and roasted limestone. Buildings are thus responsible for approximately 40% of annual greenhouse gas emissions.
In the construction of a data center, 40% is accounted for by concrete, followed by 25% of fuel, and then steel (10%). The stakeholders will need to adopt low-carbon methods to construct new data centers. This can result in a reduction of carbon emissions by 13%. Rather than building new data centers, repurposing an old building will save a lot more carbon. Otherwise known as brownfield deployment, it can also help data center companies LEED certification.
Reliance on Data Center as a Service (DCaaS) Will Increase
A 2021 survey conducted on senior IT professionals in 100 companies that invested annually $1 million in cloud computing, revealed that at 20-40%, CPU utilization is very low in these companies. These underutilized servers consume substantial energy, leading to increased costs and millions of tons of CO2 emissions.
The trend of Data Center as a Service (DCaaS) is fast gaining popularity. Here power, IT, cooling, and data center storage are standardized. Companies like Amazon and Microsoft are already providing these services with their Outposts and Azure program. DCaaS will continue to gain traction, as many other companies like HPE and Dell rather than being mere IT hardware and software suppliers, will also promote themselves as IT advisors to help enterprises run business applications in the cloud.
Datacenter capacity will continue to expand, and model-based software will make them resilient, efficient, and sustainable.
Examining the Role of AI in Data Center Efficiency
Metaverse will grab headlines through most of 2022, and so will be critical real-time decision making and computing. Under the increasingly common hybrid work model, colocation, clouds (public and private), and edge facilities, full-time management might become impossible. AI and machine learning will become important in optimizing network performances. This can also lead to a 9% operational cost savings.
In the post-pandemic world, sustainability will be at the core of global efforts. Organizations cannot fulfill this agenda without sustainable IT, and that includes data centers.